Human cervical cancer patient tissues microarray obtained 15 normal cervix specimens, 72 cervical intraepithelial neoplasia specimens, and 94 cervical cancer specimens from Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, with informed consent from all patients. All tissue specimens were confirmed by pathologist diagnosis.

Cell lines HeLa and C-4 I were preserved in Shanghai Cancer Institute, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University and cultured in DMEM medium (GIBCO) with 10% FBS and (v/v) penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂.

Immunohistochemistry staining and scoring were performed according to previous research (Yang et al., 2021b). The primary antibody used was anti-IPO7 (dilution 1:1000, ab99273, Abcam). Antibodies against Ki67 (27309-1-AP) and anti-PCNA (10205-2-AP) were purchased from Proteintech (Chicago, United States).

Total mRNA was extracted from cells using Trizol reagent (Takara) following the instructions. Quantitative real-time PCR (qRT-PCR) was performed with SYBR Green Super-mix (Takara) on ABI PRISM7500 Sequence Detection System (Applied Biosystems). Reference gene 18S was utilized to normalization. Primer sets used for IPO7 and 18s RNA examination were as follows: IPO7 forward 5′- CCCCAACACCATTATCGAGGC -3′, IPO7 reverse 5′- AGAGACTTGTGTGCTTCATTGAG -3′; 18s forward 5′-TGCGAGTACTCAACACCAACA-3′,18s reverse 5′- GCATATCTTCGGCCCACA-3′; B7H3 forward: 5′-ACAGGGC AGCCTATGACATT-3′, B7H3 reverse: 5′-GTCCTCAGCTCCT GCATTCT-3′

The formula RQ = 2−ΔCt was used to quantify gene expression levels for statistical analysis.

shRNAs against IPO7 were purchased from Gene Pharma (Shanghai, China). For IPO7 shRNA: shIPO7-1:5′- GATCCGC ATTCATCACATCATCAAACTTCAAGAGAGTTTGATGATGT GATGAATGCTTTTTTG3′; shIPO7-2:5′-GATCCGAACAGGG ATGTACCTAATGATTCAAGAGATCATTAGGTACATCCCTG TTCTTTTTTG -3′. siRNAs against GFT2I were also purchased from Gene Pharma (Shanghai, China). Transfection according to the manufacture’s protocols, using Lipofectamine 3000. For GFT2I siRNA: siGTF2I-1: 5- GCCAGAAUCACUAAAUUA -3, siGTF2I-2: 5- CAGCCACAGAAGAAAUAA -3.

Whole-cell lysates and separate nuclear/cytoplasmic fractions were extracted according to instructions and western blotting was performed as previously described (Yang et al., 2021b). Antibodies against GAPDH (60004-1-Ig), p-mTOR (67778-1-Ig), mTOR (20657-1-AP), p-AKT (66444-1-Ig), AKT (10176-2-AP), p-P70S6K (28983-1-AP), P70S6K (14485-1-AP), and GTF2I (10499-1-AP) were purchased from Proteintech (Chicago, United States). The antibodies anti-IPO7 (ab99273, Abcam).

CCK-8 cell proliferation, Colony formation, and Edu stain assay were performed as previously described (Yang et al., 2021b).

The tryptic peptides were dissolved in 0.1% formic acid (solvent A), and then separated using the EASY-nLC 1000 ultra-high performance liquid system. The peptides are separated by the ultra-high performance liquid system and injected into the NSI ion source for ionization and then analyzed by Orbitrap Fusion mass spectrometry. The ion source voltage is set to 2.2 kV, and the peptide precursor ions and their secondary fragments are detected and analyzed by high-resolution Orbitrap. The scanning range of the primary mass spectrum is set to 350–1550 m/z, and the scanning resolution is set to 60,000; the scanning range of the secondary mass spectrum is set to a fixed starting point of 100 m/z, and the secondary scanning resolution is set to 30,000. The resulting MS/MS data were processed using Maxquant search engine (v.1.5.2.8). Set the restriction enzyme digestion method to Trypsin/P; set the number of missed cleavage sites to 2; set the minimum peptide length to 7 amino acid residues. FDR was adjusted to <1% and minimum score for modified peptides was set >40.

The peptides are from the remaining peptides of proteomics. The peptides are separated by the ultra-high performance liquid system and injected into the NSI ion source for ionization and then analyzed by the Q Exactive TM Plus mass spectrometer. The ion source voltage is set to 2.2 kV, and the peptide precursor ions and their secondary fragments are detected and analyzed by high-resolution Orbitrap. Peptide parameters: The protease is set to Trypsin (KR/P), and the maximum number of missed sites is set to 0. The peptide length is set to 7–25 amino acid residues, and cysteine alkylation is set as a fixed modification.

We used the ‘‘Expression analysis-Box Plots’’ module of the GEPIA2 web server¹ to obtain box plots of the expression difference between these tumor tissues and the corresponding normal tissues of the GTEx (Genotype-Tissue Expression) database and “Match TCGA normal and GTEx data” (Tang et al., 2019). The UALCAN portal², an interactive web resource for analyzing cancer Omics data, allowed us to conduct protein expression analysis of the CPTAC (Clinical proteomic tumor analysis consortium) dataset (Chen et al., 2019). We used the ‘‘Immune-Gene’’ module of the TIMER2 web server and Sanger box³ to explore the association between IPO7 expression and immune infiltrates in cervical cancer. Meanwhile, we used the TISIDB database⁴ and verified the above results (Ru et al., 2019).

The SPSS 19.0 and GraphPad Prism 8.0 software was employed for statistical analysis. The student t-test was employed to analyze two groups of data. The one-way ANOVA was used for comprising of multiple groups. Error bars represent mean ± standard deviation (SD). Values of P < 0.05 were considered statistically significant.

To explore the potential oncogenic role of IPO7, we first applied the TCGA and GTEx datasets to analyze and results showed that the gene expression of IPO7 was significantly increased in the various cancers (Figure 1A). The CPTAC dataset also showed a higher level of IPO7 protein in the various primary tumor tissues (Figure 1B). Meanwhile, high expression of IPO7 was correlated with of poorer prognosis for cancer patients (Figure 1C). To further evaluate the expression of IPO7 in cervical cancer (CC), we initially investigated the gene expression profiles from the GSE6791 and GSE7803 datasets. These results show that IPO7 expression was significantly increased in CC compared to normal cervix tissues (Figure 1D). Moreover, higher expression of IPO7 was well correlated with a poorer prognosis of CC patients (Figure 1E). Furthermore, the immunohistochemistry staining assay in a tissue microarray (TMA) validated that IPO7 was primarily located in the nucleus of CC tissues and that the protein level of IPO7 was significantly higher in CC tissues than in the normal cervix and CIN (cervical intraepithelial neoplasia) tissues (Figure 1F). Therefore, we speculate that IPO7 has potential as an oncogenic biomarker for the progression of CC.

To investigate the biological functions of IPO7 in the proliferation of CC cells, HeLa, and C-4I, two CC cell lines with relatively higher expression levels of IPO7, were selected (Supplementary Figures 1A,B). The efficiency of IPO7 knockdown and overexpression was confirmed by quantitative real-time PCR and immunoblot analyses, respectively (Figures 2A,B). Compared to shCtrl cells, HeLa, and C-4I cells with IPO7 interference had significantly reduced cell viability (Figure 2C). Conversely, the overexpression of IPO7 significantly increased cell viability (Figure 2D). To further analyze the role of IPO7 in CC cell proliferation, we performed a clonogenicity assay and verified the above results (Figures 2E,F). To further verify the IPO7 oncogenic function in CC cell proliferation, a subcutaneous xenograft model was performed by injecting with shIPO7 in HeLa and C-4I cells. The CC tumor growth in the shIPO7 groups was remarkably inhibited, as evaluated by tumor weight and volume measurements (Figures 2G,I). Furthermore, IHC staining results revealed that the immunostaining intensities of PCNA and Ki-67, markers of cell proliferation, were significantly decreased in the shIPO7 groups (Figures 2H,J). As expected, overexpression of IPO7 promoted the tumor growth (Supplementary Figures 2A–D). Taken together, these results suggested that IPO7 was profoundly implicated in promoting CC proliferation.

Because IPO7, as a member of Kapβs, regulated cell functions mainly by transporting cargo proteins into the nucleus, the liquid chromatography–mass spectrometry (LC-MS/MS) analysis was performed to explore IPO7 cargo proteins. Expression alterations in the separate nuclear/cytosolic proteomes were analyzed upon IPO7 knockdown in HeLa cells (Supplementary Table 1). Comparative proteome analysis was conducted on three samples per group, and these differentially expressed proteins were identified with >1.5-fold changes and p-values <0.05. Comparative proteome profiling revealed hundreds of differentially expressed proteins in the shIPO7 group compared to shCtrl group (Figure 3A) and mainly concentrated in in the nucleus and cytoplasm (Figure 3B).

To fully understand the significant biological functions and pathways associated with IPO7, these differential proteins in the nucleus or cytoplasm were investigated based on bioinformatic analysis. COG/KOG categories analysis results showed that these differential proteins of cytoplasm and nucleus mainly were enriched in RNA processing and post-translational modification (Figure 3C). Further the top enriched GO and KEGG pathway terms are shown in Figure 3D. Biological process category analysis demonstrated that these proteins with enhanced localization in the cytoplasm were linked with post-translational regulation of proteins transport and RNA metabolic processes. Those proteins with reduced localization in the nucleus mainly were enriched in nucleic acid transport and gene expression. Molecular function category analysis showed that these differential proteins of cytoplasm were involved in nucleic acid binding and mRNA processing. Those differential proteins of nucleus mainly were enriched in RNA and DNA binding activity. The cellular component category indicated that these differential proteins of cytoplasm and nucleus were enriched in ribosome and organelle part. The KEGG pathway analysis revealed that the differentially expressed proteins mainly participated in RNA transport and spliceosome-associated pathways.

To further verify the nuclear/cytosolic localization of different proteins, parallel reaction monitoring (PRM) was utilized for label-free quantification by a MS. Limited by the characteristics of some proteins and the abundance of their expression, 32 potential cargo proteins were further independently validated by PRM analysis (Supplementary Table 2). Each protein was quantified with more than two unique peptides. Some proteins only identified one peptide due to sensitivity and other reasons. Overall, these PRM results mirrored the results from the nuclear/cytosolic proteome analysis-based experiment. Similar to other proteomics studies, there were some discrepancies between the two quantitative datasets (Xu et al., 2012; Wu et al., 2019). As shown in Figure 4, DNA-templated transcription-associated proteins, such as DIDO1, DNMT1, and RCOR1, mainly accumulated in the nucleus and knockdown of IPO7 significantly reduced their nuclear localization. The nuclear localization of transcription regulation-associated proteins, including PPABPN1, NCBP1, THOC3, GTF3C4, and GTF3C5, in the IPO7-knockdown group was reduced. Moreover, there were some cell division- and nucleoprotein complex localization-associated proteins, such as CKAP5, GNAI2, NMD3, and PCID2, representing the scattered localization in the nucleus and cytoplasm. PRM analysis results showed that their nuclear localization was dramatically decreased and redistributed to the cytoplasm when IPO7 was knocked down.

To further elucidate the molecular machinery correlation with IPO7 oncogenic function, we compared gene expression profiles of IPO7 high expression patients to low expression patients from TCGA datasets. Gene set enrichment analysis (GSEA) showed that the PI3K-AKT signaling pathway and mTORC1 pathway were significantly altered (Figure 5A). Previous studies demonstrated that karyopherin-mediated cargo proteins may play an important role in promoting cancer progression (Yang et al., 2021a). Then, we found that only the high expression of GTF2I and TRIP12 was associated with the poor prognosis of CC patients in 32 different proteins (Supplementary Figures 3A,B). Meanwhile, GSEA indicated that the high gene expression profiles of GTF2I were also enriched in the PI3K/AKT-mTOR signaling pathway (Figure 5B). Immunoblot analysis of separate nuclear/cytoplasmic fractions showed that the nuclear localization of GTF2I was decreased when IPO7 was knocked down (Figure 5C). Moreover, Importazole, an inhibitor of IPO7, decreased the nuclear localization of GTF2I (Figure 5D). GTF2I, as a regulator of transcription, may associates with the expression of some key genes involving in the PI3K-AKT signaling pathway. Indeed, a positive correlation of GTF2I with PIK3R1 and PIK3C2A was detected in the TCGA CECS database (Figure 5E). Moreover, we could detect a positive correlation of IPO7 with PIK3R1 and PIK3C2A, which indicated that IPO7 may contribute to regulating the PI3K/AKT-mTOR signaling pathway by mediating the nuclear import of GTF2I (Figure 5F). We further performed a western blot assay to examine the activation of PI3K/AKT signaling in shIPO7 HeLa and C-4 I cells and knockdown of IPO7 or Importazole were significantly inhibited the activation of AKT and its downstream targets, mTOR and P70S6K (Figure 5G). Moreover, immunoblot analysis showed that the activation of PI3K/Akt-mTOR was inhibited when GTF2I was knocked down in HeLa and C-4 I cells (Figure 5H). Altogether, these results demonstrated that IPO7 mediated the nuclear import of GTF2I and contributed to activating PI3K/AKT-mTOR signaling pathway.

Immune infiltration in the tumor microenvironment (TME) plays a key role in tumor progression and the efficacy of immune checkpoint immunotherapies. To further explore the impact of IPO7 in the tumor immune microenvironment, we first quantified the correlation of IPO7 and the infiltration levels of the immune cell types in CC samples by the TIMER database. As shown in Figure 6A, a negative correlation of IPO7 with CD8 T cell infiltration was detected. We found the expression of IPO7 had the strongest negative correlation with activated CD8 T cell by using Gene-Immune Analysis using Sanger box³ (Figure 6B). Consistently, we also performed a TISIDB database and verified the above results (Figure 6C). Furthermore, the immunohistochemistry staining assay in a TMA validated that the expression of IPO7 inversely correlated with CD8 T cell infiltration in CC tissues (Figures 6D,E). Numerous studies have demonstrated that immune checkpoints on cancer cell surfaces participated in tumor immune evasion. Therefore, we analyzed that the correlation of IPO7 and the expression of the immune checkpoint gene by the TIMER database. As shown in Figure 6F, a strong positive correlation of IPO7 with expression of CD276 was detected. Moreover, knockdown of IPO7 remarkedly suppressed CD276 expression in HeLa and C-4 I cells (Figure 6G). In this present study, we found that IPO7-related cargo proteins mainly were enriched in gene transcription and post-translational modification. Therefore, we measured the mRNA levels of IPO7, cargo proteins, and CD276 in the CESC TCGA database and analyzed their correlations. A positive correlation of IPO7 and cargo proteins with the expression of CD276 was detected in CC patients (Supplementary Figure 4A). Taken together, IPO7 may contribute to regulating the tumor immune microenvironment of CC.